Build fails with "socket file path is too long" when run from deep directory

[seanf]

Building on Fedora 20, I get a number of error messages like this (with long periods with no output which make me think the build is tuck):

[info] >> /home/sflaniga/NotBackedUp/src/jcabi-mysql-maven-plugin/target/mysql-dist/bin/mysqld: error while loading shared libraries: libaio.so.1: cannot open shared object file: No such file or directory
[error] #start('Config(tcpport=34756, dbuser=root, dbpassword=root, name=root, dbopts=[])', /home/sflaniga/NotBackedUp/src/jcabi-mysql-maven-plugin/target/mysql-dist, /tmp/1417590291404-0, false): thrown java.io.IOException(socket /tmp/1417590291404-0/mysql.sock is not available after 5min of waiting) out of com.jcabi.mysql.maven.plugin.Instances#waitFor[314] in 5min

I'll attach the complete log when the build finishes (or I give up waiting for it).

[yegor256]

You should run this beforehand:

sudo apt-get install libaio1

[seanf]

Thanks, seems to be yum install libaio.i686 on Fedora (I already had
/usr/lib64/libaio.so.1). I'm running the build again now...

On 3 December 2014 at 17:41, Yegor Bugayenko notifications@github.com
wrote:

You should run this beforehand:

sudo apt-get install libaio1

—
Reply to this email directly or view it on GitHub
#52 (comment)
.

[seanf]

Where can I dump a 500KB build log?

[yegor256]

Send me by email: yegor@teamed.io

[seanf]

Okay, I'll have to do that when I'm back at the office tomorrow. I think the final report said that 'parallel' failed a test, so maybe it was just the checkstyle problem ("SECOND") you fixed in master (apart from the earlier libaio problem).

[seanf]

Looks like this was the important error in the log:

[INFO] [INFO] >> 141203 19:53:17 [ERROR] The socket file path is too long (> 107): /home/sflaniga/NotBackedUp/src/jcabi-mysql-maven-plugin/target/it/parallel/target/mysql-data-first/mysql.sock

The build works when run from a shallower directory, but I think the build should avoid assuming that the current directory is a good place for a socket. See #51 and related issues.

[yegor256]

@seanf I think version 0.9 (released in #50) fixed the problem. Please confirm

[seanf]

@yegor256 Yes, moving the socket under /tmp using <socket> is working for me, thanks.